Terminal-Bench 2.1

Long-horizon command-line agent execution.

89Public tasks

86.5%Highest score

The Terminal-Bench 2.1 leaderboard

Evaluates AI agents on hard, realistic long-horizon command-line tasks — 89 curated tasks with unique environments, human-written solutions, and verification tests.

Paper Data Code Sample task

Model

Score

DeepSeek-V4.1-FlashMax

86.5%

GPT-6 AstraxHigh

85.4%

GPT-5.6 SolxHigh

84.6%

Grok 4.6xHigh

84.6%

Opus 5High

83.9%

30%

40%

50%

60%

70%

80%

90%

100%

Pass@1

Terminal-Bench 2.1

The Terminal-Bench 2.1 leaderboard

The latest on frontier AI performance, straight to your inbox.